[ Answers · AI ]

What is an AI governance framework?

By Ben Tabas, Intrica Research · Updated

Answer

An AI governance framework is the set of policies, roles, risk checks and records a company uses to decide how it builds, buys and runs AI. Two public references to build on are the NIST AI RMF, released Jan. 26, 2023, and ISO/IEC 42001, published Dec. 18, 2023. ISO/IEC 42001 is the one a company can be certified against.

The reference frameworks

  • NIST AI RMF: a voluntary US framework released Jan. 26, 2023, built on four functions, Govern, Map, Measure and Manage, per NIST's page checked Oct. 28, 2026.
  • ISO/IEC 42001: the international AI management system standard, published Dec. 18, 2023, per the IEC webstore. Accredited bodies certify companies against it.
  • OECD AI Principles: adopted in 2019 and updated in May 2024, with 47 adherents, per OECD.AI checked Oct. 28, 2026. They set five values based principles, including transparency and accountability.
  • California SB 53, signed Sept. 29, 2025, requires large frontier AI developers to publish a framework on their website, per the Governor's office.

What goes into one

Anthropic's ISO/IEC 42001 announcement, dated Jan. 13, 2025, describes its certified framework as policies for designing, building and deploying AI, testing and monitoring, transparency measures for users, and defined roles and oversight.

A startup version can be short. It needs an inventory of AI systems, an owner for each, a review before launch, monitoring after launch and a written incident plan.

How to build one

  1. List every AI system you build, buy or embed, and name an owner for each.
  2. Write a one page AI policy covering approved uses, data rules and who signs off.
  3. Add a risk and impact review to your launch checklist.
  4. Test before launch and monitor after, and log what you find.
  5. Map the result to NIST AI RMF or ISO/IEC 42001 so buyers and auditors can follow it.

See which sites AI engines cite when buyers ask about AI governance in your category.

Run the free Share of VoiceBook a call

Sources

  1. AI Risk Management Framework · NIST, Oct. 28, 2026Supports: Checked Oct. 28, 2026 for the release date, voluntary status and four functions.
  2. ISO/IEC 42001:2023 · IEC Webstore, Dec. 18, 2023Supports: Publication date and scope of the AI management system standard.
  3. OECD AI Principles · OECD.AI, Oct. 28, 2026Supports: Checked Oct. 28, 2026 for the adoption and update dates, the 47 adherents and the five values based principles.
  4. Anthropic achieves ISO 42001 certification for responsible AI · Anthropic, Jan. 13, 2025Supports: What a certified AI governance framework contains in practice.
  5. Governor Newsom signs SB 53 · Office of the Governor of California, Sept. 29, 2025Supports: SB 53's requirement that large frontier developers publish a framework.

Related questions

Intrica helps AI companies get found in Google and named in AI answers. SEO and AI search for AI companies →